Many recent visual foraging studies have focused on the cognitive processing while participants select targets within a patch: for instance, they have focused on how the perceived value of targets affects selection decisions, or how people make choices between sticking with the same target type or switching to another. We have built a generative Bayesian model (FoMo) that can make relatively accurate target-by-target predictions during foraging tasks. However, a separate strand of foraging research has focused on the question of when people terminate their search in a given area to move onto a new foraging patch. Here, we use a difficult foraging paradigm where participants collect “L-shaped” targets among Ts in either exhaustive blocks (where they must find all ten targets on each trial) or inexhaustive blocks (where they can move on at any point during a given trial but must collect a certain number of targets in total in the block). We then use the results from this paradigm to extend FoMo to allow it to make patch-leaving predictions. We compare our results to other approaches to understanding patch-leaving behaviour (such as Marginal Value Theorem) and discuss how our methods can help to bring together previously disparate parts of the visual foraging literature.